Abstract
Embedding of ultra-wideband (UWB) chips inside smartphones brands UWB technology a top contender for indoor pedestrian tracking at centimeter-level accuracy. However, the signal blocking between the wearable UWB sensor (tag) and infrastructure UWB sensors (anchors) can result in meter-level tracking inaccuracy. Accordingly, the varying nature of daily-life smartphone carrying contexts can deteriorate the UWB tracking performance depending on the degree of signal blocking by each body location or context. This paper presents the performance analysis of UWB localization accuracy under daily-life smartphone carrying contexts such as texting, calling, swinging, and pocketing. First, we implemented the extended Kalman filter-based localization algorithm using the raw UWB ranging measurements in each context. Then, we evaluated the UWB localization accuracy by conducting a realtime experiment containing four campaigns in a multipath indoor environment. In each campaign, a pedestrian is equipped with a smartphone-attached tag and a head-mounted tag for comparison. The comparative results established that the 90 th percentile of UWB localization inaccuracy increases from 0.51 to 2.72 m for smartphone-attached tag in daily-life carrying contexts than that of 0.38 m of head-mounted tag. Our investigation can contribute to the adoption of smartphone’s built-in UWB chip for indoor UWB pedestrian tracking in daily-life use cases.
Funding Information
  • Gwangju Institute of Science and Technology